This is the peer reviewed version of the following article:
Ruus, A. , Daae, I. A. and Hylland, K. (2012), Accumulation of polychlorinated biphenyls from contaminated sediment by Atlantic cod (Gadus morhua): Direct accumulation
from resuspended sediment and dietary accumulation via the polychaete Nereis virens. Environmental Toxicology and Chemistry, 31: 2472-2481,
which has been published in final form at https://doi.org/10.1002/etc.1973.
This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Use of Self-Archived Versions.
It is recommended to use the published version for citation.
Accumulation of PCBs by Atlantic Cod
1
Anders Ruus
2
Norwegian Institute for Water Research (NIVA), Gaustadalléen 21, NO-0349 Oslo, Norway
3
Phone: +47 22 18 51 00
4
Fax: +47 22 18 52 00
5
6
Total number of words: 7,213
7 8 9
ACCUMULATION OF POLYCHLORINATED BIPHENYLS FROM CONTAMINATED
10
SEDIMENT BY ATLANTIC COD (GADUS MORHUA) – DIRECT ACCUMULATION
11
FROM RESUSPENDED SEDIMENT AND DIETARY ACCUMULATION VIA THE
12
POLYCHAETE NEREIS VIRENS
13 14
Anders Ruus,*† Ingrid Aarre Daae, ‡ Ketil Hylland, †‡
15
† Norwegian Institute for Water Research, Gaustadalléen 21, NO-0349 Oslo, Norway
16
‡ University of Oslo, Department of Biology, PO Box 1066, Blindern, N-0316 Oslo, Norway
17 18 19
* To whom correspondence may be addressed ([email protected]).
20 21 22
Abstract
23 24
Bioaccumulation of sediment associated polychlorinated biphenyls (PCBs) was examined in
25
Atlantic cod (Gadus morhua) through (1.) direct diffusion from the sediment (via the water
26
phase), and (2.) through the food chain (dietary exposure). To facilitate direct accumulation from
27
the sediment, the sediment was continuously resuspended. To study the dietary bioaccumulation
28
of PCBs, cod were fed benthic polychaetes (Nereis virens) previously exposed to test sediments,
29
i.e. “naturally” polluted sediments from the inner Oslofjord (Norway). Both exposure
30
experiments had duration of 129 days. Furthermore, the role of sediments as source of PCBs
31
accumulated in Oslofjord cod was elucidated, using results from environmental monitoring as a
32
reference. Generally, the results suggest that the contaminated sediments of the inner Oslofjord
33
are an important source of legacy PCBs for accumulation in resident cod, although additional
34
contributions also may be important. Crude estimates of assimilation efficiency of ingested PCBs
35
(through diet) was found to be 30-50%; highest for the lower chlorinated congeners (PCB-28 and
36
-52). Challenges for applying Trophic Magnification Factors (TMF) for determining
37
biomagnification in laboratory experiments, in terms of preventive environmental safety, are
38
indicated. The results provide useful information for parameterization of models describing the
39
behaviour of hydrophobic persistent contaminants in the foodweb of the Oslofjord and elsewhere.
40 41 42
Key Words: Bioaccumulation, PCB, Gadus morhua, sediment, Nereis virens
43 44
45
Introduction
46 47
Polychlorinated biphenyls (PCBs) and bioaccumulation processes
48
The identification of polychlorinated biphenyls (PCBs) in samples of biota by Søren Jensen in
49
the 1960s [1] initiated extensive investigation on their abundance in the environment, and their
50
distribution throughout the biosphere is now well documented [e.g. 2-4]. The banning of PCBs in
51
several countries was to follow in the 1970s and caused the global PCB production to decline.
52
One important international agreement in this regard is the Stockholm Convention on persistent
53
Organic Pollutants (POPs), which is a global treaty to protect human health and the environment
54
from hazardous substances by restricting and ultimately eliminating their use, trade, release and
55
storage. Worldwide, significant quantities of PCBs are however still in present in old
56
infrastructure and equipment. Some PCBs are shown to have various toxic effects (Reviewed by
57
Safe [5]), including immunosuppressive and endocrine disrupting effects, as well as impairment
58
of reproduction.
59 60
The environmental fate of contaminants, such as PCBs, is an important ecotoxicological
61
aspect, and bioaccumulation is a fundamental phenomenon in this regard. For a chemical to
62
bioaccumulate, it must be available (bioavailable), and once bioaccumulated, a contaminant may
63
(dependent on its physico-chemical properties) be further subject to biomagnification (the
64
chemical concentration in an organism exceeds that in its diet after dietary absorption [6]). In
65
aquatic organisms, bioaccumulation is the process that causes an increased chemical
66
concentration in the organism compared to that in its ambient environment, water and/or
67
sediment [7]. Recently a group of experts has suggested the following definition of a
68
bioaccumulative substance in a regulatory context: a substance is considered bioaccumulative if it
69
biomagnifies in food chains [8].
70 71
It is well known that because of their persistence and lipophilicity, PCBs have the potential to
72
bioaccumulate and biomagnify in food chains . The highest concentrations of these compounds
73
are found in top predators like seabirds and marine mammals [e.g. 3, 4].
74 75
Other persistent organic pollutants (POPs) share similar physicochemical properties as some
76
of the PCBs (for instance polybrominated diphenyl ethers, PBDEs and hexachlorocyclododecane,
77
HBCD; [9, 10]). Therefore, results obtained from bioaccumulation studies where PCBs are
78
employed as the model compounds may to some extent serve as valuable information with regard
79
to POP bioaccumulation processes, in general.
80 81
Aquatic organisms take up PCBs and other lipophilic substances through the ingestion of food
82
and directly from water through passive diffusion at the body surface, mainly via the respiratory
83
surfaces. Several models have been introduced to describe these processes (reviewed by Mackay
84
and Fraser [7]). Which of these routes that are the most important for bioaccumulation may vary
85
between organisms with different modes of living, and have been the subject of much discussion
86
(See below; [e.g. 7, 11, 12]). Bioaccumulation is the net result of uptake and elimination (the
87
latter through metabolic transformation, reproductive losses, fecal egestion, or diffusive fluxes
88
[13, 14]). The capability of metabolic transformation of PCBs by fish is however limited, and
89
fecal elimination has been shown as a no important loss mechanism [13]. Mechanistic mass
90
balance models may be built where the different uptake and elimination processes are quantified.
91
These models have the advantage that they may take into account effects of phenomena like
92
compound specific biotransformation rates and ‘growth dilution’ [7]. They are, however, in need
93
of sound parameterization.
94 95
Environmental monitoring
96
The Coordinated Environmental Monitoring Program (CEMP) is administered by the Oslo
97
and Paris Commissions (OSPAR) in their effort to assess and remedy anthropogenic impact on
98
the marine environment of the North East Atlantic. The Norwegian contribution to the CEMP
99
was initiated by the Norwegian Climate and Pollution Agency in 1981 as part of the national
100
monitoring program, and the current focus is on the levels, trends and effects of hazardous
101
substances, including PCBs. It comprises several areas, including the Oslofjord and adjacent
102
localities [15].
103 104
Objectives
105
The objective of this study was to elucidate the role of sediments as source of PCBs
106
accumulated in Atlantic cod (Gadus morhua) through two exposure routes: (1) through (direct)
107
diffusion from the sediment (via the water phase), and (2) through the food chain (dietary
108
exposure). Furthermore, known PCB-concentrations in liver of cod from the inner Oslofjord,
109
available through a national environmental monitoring program (CEMP; described above), were
110
used as reference to assess the role of contaminated sediments specifically for the cod in the inner
111
Oslofjord.
112 113
Current chemical legislation and regulating organs use a framework and criteria to assess the
114
potential hazard and risk according to the chemicals’ bioaccumulative potential (B), in addition to
115
their persistence (P) and toxicity (T) (“PBT” criteria; [e.g. 16]). These criteria are continuously
116
challenged [e.g. 8, 17]. Based on recent discussions among scientists and regulators, several
117
recommendations have been put forward regarding evaluation of the B-criterion [e.g. 8]. These
118
recommendations include taking into account the accumulation from the diet by the use of
119
biomagnification factors (BMF; ratio between predator and prey concentrations) and/or trophic
120
magnification factors (TMF; the average factor by which the lipid normalized concentration
121
increases per trophic level; determined from the slope (m) derived by linear regression of Log10-
122
transformed biota concentration and trophic position; TMF = 10m) when evaluating the
123
bioaccumulation potential of a chemical. The present study also serves as a trial for the feasibility
124
of such an approach.
125 126
As such, organisms used in the present study were Atlantic cod and the “King rag” worm
127
Nereis virens (Polychaeta). The study has comprised two long term (months) mesocosm
128
experiments:
129
1. Study of the bioaccumulation of PCBs in cod exposed to resuspended contaminated
130
sediment particles (‘the sediment resuspension experiment’).
131
2. Study of the bioaccumulation of PCBs in cod fed benthic invertebrates (the polychaete
132
Nereis virens) exposed to contaminated sediment (‘the dietary exposure experiment’).
133 134
In both exposure experiments, cod were exposed for a total of 129 days, with sampling at d 0,
135
d 13, d 26, d 39, d 52, d 66, d 97 and d 129. In the latter experiment, the polychaetes were
136
exposed to sediment for a minimum of 9 weeks before being fed to the cod.
137 138
The organisms employed were chosen for the commercial value, ecological relevance, the
139
availability, and the experience that they are possible to hold in aquaria for extended periods.
140
Furthermore, Atlantic cod is also one of the species of choice in several environmental
141
monitoring programs, including CEMP. The cod is common on the continental shelf in most of
142
the North-Atlantic. Mostly, the cod is a benthic feeder, but may live pelagic. Nereis virens is
143
common along the Atlantic coasts of Europe, North to the mid-West coast of Norway [18]. It
144
occupies burrows in muddy sand. Sediment-dwelling organisms, such as several species in the
145
Nereis genus are important prey items e.g. to demersal and bottom-feeding fish, such as cod, and
146
may therefore contribute to the transport of contaminants to higher levels in marine food chains
147
[e.g. 4].
148 149
The contaminated sediments employed in the experiments were from the inner Oslofjord,
150
which includes the city harbor area of Oslo. The Norwegian Food Safety Authority has issued
151
advice against consumption of cod liver from the inner Oslofjord, based on the PCB
152
contamination.
153 154
In the present experiments, samples were also preserved for the evaluation of metabolites of
155
polycyclic aromatic hydrocarbon (PAHs) in the bile of the fish, as well as for different biomarker
156
responses. These will be discussed elsewhere (Daae et al. in prep.).
157 158
Materials and methods
159
Sediment sampling
160
The test-sediment (PCB-contaminated) was collected from the upper 5-15 cm of the sediments
161
at locations in the Inner Oslofjord area (Eastern Norway), using a 0.1 m2 Van Veen grab. The
162
collection took place between 59 52.176 and 59 53.974 North and between 10 40.630 and
163
10 43.682 East. Uncontaminated reference (control) sediment was collected at a fixed location
164
in the outer Oslofjord, previously employed in bioaccumulation studies and documented to have
165
very low concentrations of organic pollutants [19]. The sediments were collected in spring, 2006.
166 167
For transport and prior to the experiments, the sediment was stored in 150-L boxes.
168
Approximately 750 L of contaminated sediment (6 boxes) and 250 L of reference (control)
169
sediment (2 boxes) were collected. The sediment was homogenized by shoveling aliquots of
170
sediment between boxes simultaneously as they were slurried by the use of a mortar mixer for
171
approximately 1 h (Eibenstock EHR-20 S, Elektrowerkzeuge GmBH Eibenstock, Germany).
172 173
Test-organisms
174
Atlantic cod were purchased from Marin Invest AS (Sandøy, Western Norway; resuspended
175
sediment exposure experiment) and Marine Harvest ASA (Eggesbønes, Western Norway; dietary
176
exposure experiment). The fish were brought to NIVA’s marine research facility Solbergstrand
177
by the use of tank lorries and held for a minimum of 2 months (acclimation) before initiation of
178
the experiments. Prior to arrival, the fish were fed pellets: Gemma micro, Gemma 0.3/0.5,
179
Gemma 0.75/1.0/1.2, Europa Respons 1.5 mm, Europa Respons 2.0 mm and Europa Respons 3.0
180
mm. After arrival at Solbergstrand, prior to the experiment, fish were fed Europa Respons 3.0 and
181
4.0 mm (supplier of all fish feed; Skretting AS, Stavanger, Norway). The experiments were
182
conducted after approval by The Norwegian Animal Research Authority (NARA).
183 184
Rag worms (Nereis virens) were purchased from Seabait Ltd. (Ashington Northumberland,
185
UK), and brought to NIVA’s marine research facility Solbergstrand by air freight and car. Before
186
and during the experiments, the worms were fed Skretting advanced fish feed (Coapse fish - 23.
187
Skretting, Roman Island, Westfort Co., Mayo, Ireland).
188 189
Experimental setup and sampling procedures
190
The experimental procedures for ‘the sediment resuspension experiment’ were as follows:
191
Atlantic cod (approximately 450 g) were transferred to 6 fiberglass tanks (45 110 110 cm;
192
545 L) of which 3 tanks (the ‘exposed’ group) contained a 16 cm deep layer of sediment from the
193
inner Oslofjord (approximately 195 L of sediment in each tank; samples recovered for chemical
194
analysis). The remaining three tanks did not contain sediment (‘control’ group). At day zero (d 0;
195
March 3rd, 2006) 13 individual cod were transferred to each tank.
196 197
The tanks were supplied with running seawater (8 L min-1; from 60 m depth outside the
198
research facility Solbergstrand). In this way the fish were ensured sufficient oxygen (measured to
199
75% saturation; WTW Oxi 340i; WTW GmbH, Weilheim, Germany). Through the exposure
200
period (129 days) the mean temperature was 7.4 C (range: 6.3-9.2) and the mean salinity was
201
34.6 (range: 34.2-34.9; logged by WTW-probes, WTW GmbH). The fish were given a
202
maintenance diet (every second day) of pellets (3 mm and 4 mm; sampled for chemical analysis)
203
throughout the experiment to comply with their needs, but avoid excessive growth. Because of
204
the proportion of sediment in relation to amount water and fish, the swimming activity of the fish
205
could initially disturb the sediment sufficiently to produce turbid water. Mechanical disturbance
206
of the sediment was performed the last 4-5 weeks by the use of a small propeller (3 blades; : 4
207
cm) mounted on a drill (Bosch P9B 600 RE; Robert Bosch AS, Ski, Norway). Sampling of fish
208
were performed at d 0, d 13, d 26, d 39, d 52, d 66, d 97 and d 129. Six fish were sampled at day
209
0. At every other outtake, one fish from each tank were sampled (n=3 in each group, ‘exposed’
210
and ‘control’). The fish were terminated by a blow to the head, before the gall-bladder was
211
emptied of bile (using a syringe; handled elsewhere (Daae et al., in prep.)) and the liver was
212
carefully excised and stored for chemical analysis (-20 °C; cod is a lean fish with the liver as the
213
storage site for lipid reserves, thus nearly the whole body burden of lipophilic contaminants can
214
be observed here [15]).
215 216
The experimental procedures for ‘the dietary exposure experiment’ were carried out in two
217
phases, (1.) exposure of polychaetes to sediments and (2.) feeding polycheates to fish:
218 219
The exposure of polychaete worms was as follows: N. virens were exposed to the sediments
220
(inner Oslofjord (‘exposed’) or outer Oslofjord (‘control’)) in containers of 11 L with lid.
221
Approximately 8 L of sediments and 20-35 worms were added to each container, which was
222
supplied with continuous water flow through (250 mL min-1). One container was prepared for
223
each feeding of fish (a total of 37 feedings). For logistical reasons, two rounds of polychaete
224
exposure were conducted. Worms from the first exposure, were individually stored at -20 C and
225
served as ‘box lunch‘ for the fish towards the end of the fish exposure period (last 3 weeks).
226
Furthermore, this batch functioned as the food backup, in case of unexpected mortality among the
227
worms in the second batch. The worms from the second batch were extracted fresh from the
228
sediment prior to each feeding of fish. Triplicate samples were prepared of sediments and
229
polychaetes for chemical analysis.
230 231
The worms were fed pellets (see above, 2-3 g per container) 3 times each week, and were
232
exposed to the sediments for a minimum of 9 weeks (which is twice the minimum duration
233
recommended by Lee et al. [20]). Through the polychaete exposure periods the mean
234
temperatures were 8.1 C (range: 7.6-9.2) and 8.5 C (range: 5.8-12.1), while the mean salinities
235
were 34.3 (range: 33.9-34.5) and 34.1 (range: 33.4-34.5) for batch 1 and 2, respectively (logged
236
by WTW-probes, WTW GmbH).
237 238
The feeding of sediment exposed-polychaetes to Atlantic cod was as follows: One week prior
239
to the first feeding (d 0) the cod (mean weight: 78 g) were transferred to individual compartments
240
in aquaria measuring 35 35 70 cm (3 compartments in each). One fish was added to each
241
compartment. A total of 54 fish were thus occupying 18 aquaria. The aquaria were supplied air
242
(bubbling) and continuous water flow through (1 L min-1). Through the exposure period (129
243
days) the mean temperature was 7.7 C (range: 6.6-9.7) and the mean salinity was 34.3 (range:
244
33.8-34.5; logged by WTW-probes, WTW GmbH).
245 246
The cod were fed exclusively N. virens twice a week (every 3rd to 4th day). The amount of
247
worm (4-6 g) fed to the fish was weighed out and logged. The weekly amount of worm fed to the
248
fish represented a minimum of 8% of the fish body weight. The individual compartments in the
249
aquaria facilitated the individual feeding of the fish and at each feeding it was observed that the
250
fish ingested all that was presented.
251 252
Sampling of fish was performed at d 0, d 13, d 26, d 39, d 52, d 66, d 97 and d 129. At d 0, six
253
fish were sampled. At every other outtake, 3 fish were sampled from each group (fed worms
254
exposed to contaminated sediment (‘exposed’) or fed worms exposed to clean sediment
255
(‘control’)). The fish were put to death by a blow to the head. At each sampling the fish length,
256
weight and liver weight were measured. Samples were secured from the liver and stored (-20 C)
257
for chemical analysis. Furthermore, samples were preserved from bile, liver and blood for
258
analysis of metabolites of polycyclic aromatic hydrocarbons (PAHs; in bile), activity of 7-
259
ethoxyresorufin O-deetylase (EROD; in liver), amount of cytochrome P450 1A protein (CYP1A;
260
in liver), amount of vitellogenin and zona radiata protein (in blood), and activity of -amino
261
levulinic acid dehydrase (Ala-D; in blood). These biomarker responses are handled elsewhere
262
(Daae et al., in prep.).
263 264
Extraction, cleanup and PCB analysis, and analysis of sediment properties
265
The chemical analyses were performed at NIVA. The laboratory is accredited by the
266
Norwegian Accreditation as a testing laboratory according to the requirements of NS-EN
267
ISO/IEC 17025 (2000). Furthermore, analytical standards are certified by the participation in
268
international calibration tests, including QUASIMEME twice per year. The procedures for
269
extraction, cleanup and quantification of PCB congeners were as described in Supplemental
270
information, as are the analyses of sediment properties. The certified reference materials used
271
were SRM 1944 and SRM 1588b (National Institute of Standards and Technology, Gaithersburg,
272
MD, USA) and recoveries were 78 to 120 %. The detection limit was defined as >3 times signal
273
noise and was from <0.05 to <1.0, dependent on congener and matrix.
274 275
Statistical methods
276
Statistical analysis was performed with the use of Statistica software (Ver 7.0;
277
Statsoft,Tulsa, OK, USA). Temporal differences in cod liver PCB concentrations (within groups;
278
“exposed” or “control”) were evaluated using Analysis of Variance (ANOVA). Levene’s test was
279
used to test for heterogeneity of variance. If necessary, data were Log10-transformed to reduce
280
heterogeneity of variance. Furthermore, if homogeneity of variance was not obtained, temporal
281
differences in PCB concentrations were evaluated using the non-parametric Kruskal-Wallis test,
282
as were differences in PCB concentrations between cod exposed to contaminated sediment
283
(directly or via polychaetes) and unexposed cod (no sediment exposure, or fed polychaetes
284
exposed to clean control sediment), and differences in PCB concentrations between polychaetes
285
exposed to contaminated sediments and polychaetes exposed to clean (control) sediments. The
286
Dunnet post-hoc test (following ANOVA), or the non-parametric multiple comparison test
287
(following Kruskal-Wallis), were employed to test for differences against zero-time. Linear
288
regressions were applied to assess concentration increases in cod. A significance level of =
289
0.05 was chosen.
290 291
Results and Discussion
292
Methodical aspects
293
There was no mortality of cod during the exposure experiments, apart from one individual in
294
the dietary exposure experiment (a surplus of fish was employed in the experiments (see above),
295
thus this had no effect on the number of analyzed individuals). Apparently there was no, or
296
minimal (not logged) mortality among the worms during the exposure, as there were plenty of
297
worms in surplus for the feeding of cod, and no cadavers could be observed. The cod from the
298
dietary exposure experiment showed no signs of discomfort from a diet consisting exclusively of
299
polychaetes. They soon became very tame, eating from the hand of the keeper. Furthermore, by
300
day 129 of the exposure, they had gained 46.5% (mean ± 7.6 standard deviation) of their initial
301
bodyweight (measured at d 0; corresponding to 33 g from a starting point of 71 g, on average),
302
indicating that they were thriving on the worms. The holding of the fish, however (in terms of
303
size of the setup) dictated limitations in the number of replicates (n=3).
304 305
The sediments applied in the two exposure experiments differed somewhat in PCB-content
306
(see below; Table 1), despite the homogenization efforts (above). This renders direct comparisons
307
between absolute concentrations accumulated in the fish from the two exposure experiments
308
difficult. It should be noted, however, that the variability between replicates, within each
309
experiment, was small. Direct comparisons between absolute concentrations accumulated in the
310
fish from the two exposure experiments were further complicated by different lipid content (and
311
different variability in such) of the fish livers, between exposure experiments (see below; Figure
312
1; Table S1, see Supplemental information).
313 314
It should also be noted that the because of the fairly high water flow-through (to meet the life
315
support requirements of the fish) in the ‘sediment resuspension experiment’, the PCB distribution
316
in the exposure system may not reflect partition equilibrium between sediment and water [21].
317
This may obscure the importance of PCB accumulation from sediment via the water phase.
318
However, the flow-through conditions will resemble field conditions, where mixing and water
319
movements will be present. On the other hand, resuspension of the sediment (to mimic
320
disturbance of sediment in shallow waters) was done to facilitate desorption of particle associated
321
PCBs and render them more available to the fish.
322
323
Sediments and polychaetes
324
Moderately high concentrations of PCBs were observed in the sediments used in the
325
experiments (Table 1; [22]), with concentrations a factor of ~4 higher in the dietary exposure
326
experiment than in the sediment resuspension experiment.
327 328
Concentrations of PCBs accumulated in N. virens were significantly higher in the exposed
329
worms than in the control group (a factor of 3 to 6; Table 1). The lipid content in the worms was
330
identical between groups. Calculating biota-to-sediment accumulation factors (BSAF;
331
(COrg/fLip)/(CSed/fOC), where COrg is the wet wt. concentration in the organism, fLip is the fraction of
332
tissue wet wt. that is lipid, CSed is the dry wt. concentration in the sediment, and fOC is the fraction
333
of organic carbon in the sediment (g g-1 dry wt.)) gave values between 0.24 (PCB-28) and 0.67
334
(PCB-101). These values are somewhat lower (implying lower bioavailability) than a theoretical
335
expectation of 1.6 (see Supplemental information), provided the following assumptions [23]: (1.)
336
bioaccumulation of sediment associated PCBs in N. virens occurs (merely) as an equilibrium
337
partitioning between sediment particles (organic carbon in particular) and water, and between
338
water and the organism lipids, (2.) the relationship between the sediment:water partition
339
coefficient (Kd) and the organic carbon:water partition coefficient (KOC) is Kd = KOC fOC, (3.)
340
There is a double logarithmic, linear relationship between KOC and KOW (the octanol:water
341
partition coefficient; log KOC = log KOW – 0.21; [24]; one domain sorption model) , and (4.) the
342
partitioning coefficient between the organism lipids and the water equals KOW. Furthermore,
343
BSAFs of PCBs were somewhat lower than those e.g. observed in the oligochaete Lumbriculus
344
variegatus [25, 26]. On the other hand, BSAFs were orders of magnitude higher than those
345
observed for polycyclic aromatic hydrocarbons (PAHs) in e.g. N. diversicolor exposed to
346
sediments with characteristic composition of sorption domains with high binding strength [23].
347
The values corresponded, however, well with previously observed BSAFs for PCBs in N.
348
diversicolor [19] and grass shrimp (Palaemonetes pugio; [27]). The results indicate fairly high
349
bioavailability of PCBs in the sediments, possibly slightly reduced by carbonaceous geosorbents
350
present in the Oslofjord sediments [28].
351 352 353 Cod
Different lipid content in fish livers were (as mentioned) observed between exposure
354
experiments (Figure 1; Table S1, see Supplemental information). Furthermore, the variability in
355
lipid content among livers were different between exposure experiments (coefficient of variation,
356
CV = 20.3% and 12.8% in the dietary exposure experiment and the sediment resuspension
357
experiment, respectively; all individuals and sampling days). There were, however, no signs of a
358
systematic change in lipid content, over time, in neither of the experiments, or groups (exposed
359
vs. control); Figure 1; Table S1, see Supplemental information). Consequently, concentrations are
360
treated/graphically expressed on a lipid wt. basis in the following (wet wt. concentrations
361
presented in Table S1; see Supplemental information).
362 363
PCBs and other hydrophobic compounds express a high affinity for lipids [e.g. 7].
364
Ideally, equilibrium will eventually occur between the concentrations of these compounds in
365
aquatic organisms and the surrounding water constituting their habitat [12]. Respiratory surfaces
366
(i.e. gills) play an important role in this partitioning, as the compounds associate with the lipoid
367
cell membranes in the gill epithelium and are circulated to lipid tissues within the organism.
368
Equilibrium partitioning can be regarded as an approximate lipid:water partitioning, thus the KOW
369
may provide valuable information [7]. The PCB congeners in focus of the present study have
370
KOW values ranging from 5.13 103 (PCB-28) to 1.54 107 (PCB-180), increasing with degree of
371
chlorination [29].
372 373
An apparent increase in concentrations with time could be observed in the exposed group of
374
the sediment resuspension experiment for most congeners (Figure 2). However, the hepatic
375
concentrations of several congeners apparently also increased towards the end of the experiment
376
in the control group (Figure 2). Nevertheless, significant differences were found between the
377
exposed group and the control group, at several sampling days, but only for PCB-28 and -52
378
(those with the lowest KOW; note limited statistical power due to low n). Furthermore,
379
significantly different concentrations towards the end of the experiment, compared to d 0, were
380
found for these congeners. The apparent increase, also in the control group, may likely be
381
explained by fish in both groups being fed with commercial fish feed throughout the experiment.
382
Analysis of this feed showed traces of PCBs (0.25 µg kg-1 (PCB-28 and -180) to 1.7 µg kg-1
383
(PCB-153) wet wt.; PCB7=6.75 µg kg-1 wet wt.; lipid content 16.0% wet wt.).
384 385
Ergo, the two congeners with the lowest hydrophobicity (KOW) showed a temporal increase in
386
concentrations, that may be related to accumulation of sediment associated PCBs, corresponding
387
to previous observations [e.g. 27], suggesting lower bioavailability of higher chlorinated
388
congeners in the water phase. According to Clark et al. [11], a large fraction of chemicals with
389
KOW 104 – 105 may be present in the water phase (dissolved), when KOW=106, half is adsorbed to
390
particles present in the water, and when KOW=108, all is adsorbed to particles. Furthermore,
391
several field observations suggest that aquatic organisms that accumulate PCBs from water
392
(through diffusion), contain higher proportions of the lower chlorinated congeners [e.g. 3, 4].
393 394
The results further suggest that steady state is not reached (no indication of an asymptotic
395
levelling) after 129 days for any of the congeners. Congeners with a lower degree of chlorination
396
(and thus lower hydrophobicity) reach equilibrium faster than the higher chlorinated homologues
397
[e.g. 30, 31]. An influence on the results by congener specific biotransformation by the fish can,
398
however, not be ruled out.
399 400
There were markedly (statistically significant) higher concentrations of all PCB congeners in
401
the exposed group, compared to the control, towards the end of the dietary exposure experiment
402
(Figure 3). The PCB concentrations in the unexposed (control) group maintained a low level
403
through the whole experiment (129 days; Figure 3). Significant differences in concentrations
404
among sampling days and compared to d 0 could be observed (again) for congeners PCB-28 and
405
-52 (significant differences among sampling days in the exposed group also for PCB-138 and -
406
180; Figure 3; note low statistical power due to low n). Also in the dietary exposure experiment,
407
there were no indications of an asymptotic levelling of the concentrations within the maximum
408
exposure period of 129 days (Figure 3). Thus concentrations might very well have increased if
409
the experiment was continued. This possible continued increase also illustrates challenges using
410
biomagnification as a regulatory endpoint [8], if such potential must be shown prior to chemicals
411
being released to the market and thus the environment (e.g. according to the Registration,
412
Evaluation, Authorisation and Restriction of Chemicals (REACH) of the European Union [16]).
413
The Trophic Magnification Factor (TMF) is suggested as a “golden standard” in bioaccumulation
414
and has been applied in many field studies [e.g. 8]. The present accumulation results, however,
415
suggests inappropriately complex, time consuming and expensive test protocols if TMFs would
416
be applied to laboratory experiments, in terms of preventive environmental safety. Thus, the use
417
of alternative approaches, such as measuring uptake and elimination rates (in an uptake phase and
418
a subsequent depuration phase), to derive “steady-state biomagnification factors” [e.g. 8] seems
419
more applicable in this regard.
420 421
Crude estimates of the assimilation efficiency of the PCBs fed to cod, through the polychaete
422
“vehicle”, during the 129 d exposure period could be made since the following parameters were
423
known: (1.) the total amount (kg) polychaetes fed to the cod (2.) the mean PCB concentrations
424
(µg kg-1) in the polychaetes, (3.) initial (d 0) PCB concentrations (µg kg-1) and weight (kg) of cod
425
livers, (4.) terminal PCB concentrations (µg kg-1) and weight (kg) of cod livers. The results show
426
that 30-50% of the total amount of PCBs fed to the cod (via N. virens) through the 129 d
427
exposure period is stored in the cod liver (Table S2; see Supplemental information). The highest
428
assimilation efficiency was apparent for the lower chlorinated congeners (PCB-28 and -52).
429 430
According to Kelly et al. [32], the assimilation efficiency of different persistent organic
431
compounds in fish is slightly less than 50% and decrease for compounds with KOW>107. It is
432
suggested that transport of very hydrophobic compounds across the intestinal wall is limited by
433
an aqueous diffusion resistance [33]. Thus, a possible explanation for the decrease in dietary
434
assimilation efficiency with increasing hydrophobicity, is slow transport through intestinal
435
aqueous phases because of low aqueous solubility [34, 35]. An influence on the results by
436
congener specific biotransformation by the fish can, however, not be ruled out.
437 438
As mentioned, there are factors that impede direct comparisons between the results of the
439
sediment resuspension experiment and the dietary exposure experiment. Firstly, the sediment
440
applied in the dietary exposure experiment contained somewhat higher concentrations of PCBs,
441
than the sediment applied in the sediment resuspension experiment (Table 1). Secondly, there
442
were differences in the liver lipid content of the fish employed in the two experiments (Figure 1;
443
Table S1, see Supplemental information). In a review of bioaccumulation mechanisms and
444
models, Mackay and Fraser [7] present a “rule of thumb” implying that the importance of dietary
445
accumulation versus diffusive accumulation (across respiratory surfaces) is approximately
446
KOW/200 000. This relationship will vary dependent on fish size, condition and species. However,
447
for very hydrophobic substances (i.e. log KOW>6.5) diffusive uptake over respiratory surfaces will
448
not be important, while for less hydrophobic substances (i.e. log KOW<4.0), dietary uptake
449
becomes less important, since equilibrium between the fish and the surrounding water will be
450
reached more quickly. The results of the present study (considering the above mentioned
451
complicating factors, however) do not suggest this “rule of thumb” erroneous.
452 453
Extrapolations and concluding remarks
454
In the dietary exposure experiment, higher concentrations were observed in the exposed
455
group, compared to the control towards the end of the exposure period (d 52 – d 129) for all
456
congeners (Figure 3). Furthermore, no increases in concentrations were indicated in the control
457
group (Figure 3). Plotting time (days; continuous scale) versus concentration (exposed group),
458
produced significant (p<0.0014) linear regressions for all congeners (as well as PCB7; Figure
459
S1, see Supplemental information). The goodness-of-fit decreased, however, for the more
460
chlorinated/hydrophobic congeners (R2= 0.76, 0.68, 0.40, 0.39, 0.34, 0.44, 0.44 and 0.43 for
461
PCB-28, -52, -101, -118, -153, -138, -180 and PCB7, respectively; Figure S1, see Supplemental
462
information). Given the following assumptions: (1.) a continued linear increase in concentrations
463
and (2.) an initial concentration equal to the intercept of the regression (approximately the
464
medians of the d 0 concentrations; see Figure S1, Supplemental information), the slopes of the
465
regressions may be used to make crude estimates/extrapolations of the time needed to reach
466
concentrations present in wild caught cod from the inner Oslofjord (known through
467
environmental monitoring; Table 2). Such extrapolations showed that the time needed to reach
468
concentrations present in wild Oslofjord cod were 0.2 (PCB-28) to 5.8 (PCB-153) years (Table
469
3). It must be noted that these extrapolations may likely represent underestimates, since the
470
assumption of a continued linear increase until reaching concentrations present in wild Oslofjord
471
cod might be erroneous. Alternatively, the increase might be curvilinear (first order; [e.g. 36,
472
37]). Additionally, the issue of growth dilution must be taken into account. For compounds with
473
concentrations that change slowly, a growth constant of e.g. 0.001 Day-1 (corresponding to a
474
doubling in size in slightly less than 2 years) will lead to a considerable dilution in the organism
475
[7]. Other factors will also increase the uncertainty of such crude extrapolations. Wild cod also
476
feed on other organisms than polychaetes [e.g. 38], and at a certain size, a shift in trophic position
477
may occur. Furthermore, the PCB concentrations of the Oslofjord sediment are obviously not
478
uniform [e.g. 39] and will be both higher and lower than those used in the experiment in some
479
areas. Nevertheless, generally the results suggest that the contaminated sediments of the inner
480
Oslofjord are an important source of legacy PCBs for accumulation in the native cod, although
481
additional contributions from e.g. atmospheric deposition and runoff from the surrounding
482
(urban) landscapes also may be substantial [40]. The study has further indicated the feasibility of
483
conducting long term (months) experiments for elucidating contaminant accumulation from
484
sediments to fish, via one level of the food chain, providing opportunities for related topics. On
485
the other hand, challenges for applying Trophic Magnification Factors (TMF) to determine
486
biomagnification in laboratory experiments, in terms of preventive environmental safety, are
487
indicated. The results will provide useful information for parameterization of models describing
488
the behaviour of hydrophobic persistent contaminants in the foodweb of the Oslofjord and
489
elsewhere.
490 491 492
Supplemental information
493
Extraction, cleanup and PCB analysis, Sediment property analyses, Table S1, Table S2, Figure
494
S1, Calculation of biota-to-sediment accumulation factors (BSAFs).
495 496 497
Acknowledgements
498
This study was partly (50%) funded by “Fagrådet for vann- og avløpsteknisk samarbeid i indre
499
Oslofjord”. Thanks to Jan Magnusson for assistance in launching the project. Thanks are also due
500
to Sigurd Øxnevad, Per-Ivar Johannessen and Nasir Hamndan El-Shaikh for their skillful
501
assistance during the mesocosm exposure experiments.
502 503 504
References
505 506
[1] Anon. 1966. Report of a new chemical hazard. New Scientist 32:612.
507
[2] MacDonald RW, Barrie LA, Bidleman TF, Diamond ML, Gregor DJ, Semkin RG,
508
Strachan WMJ, Li YF, Wania F, Alaee M, Alexeeva LB, Backus SM, Bailey R, Bewers JM,
509
Gobeil C, Halsall CJ, Harner T, Hoff JT, Jantunen LMM, Lockhart WL, Mackay D, Muir DCG,
510
Pudykiewicz J, Reimer KJ, Smith JN, Stern GA, Schroeder WH, Wagemann R, Yunker MB.
511
2000. Contaminants in the Canadian Arctic: 5 years of progress in understanding sources,
512
occurrence and pathways. Sci Tot Environ 254:93-234.
513
[3] Ruus A, Ugland KI, Espeland O, Skaare JU. 1999. Organochlorine contaminants in a
514
local marine food chain from Jarfjord, Northern Norway. Mar Environ Res 48:131-146.
515
[4] Ruus A, Ugland KI, Skaare JU. 2002. Influence of trophic position on organochlorine
516
concentrations and compositional patterns in a marine food web. Environ Toxicol Chem 21:2356-
517
2364.
518
[5] Safe SH. 1994. Polychlorinated biphenyls (PCBS) – Environmental impact, biochemical
519
and toxic responses, and implications for risk assessment. Crit Rev Toxicol 24:87-149.
520
[6] Gobas F, Morrison H. 2000. Bioconcentration and biomagnification in the aquatic
521
environment. In Boethling R, Mackay D, eds, Handbook of property estimation methods for
522
chemicals: environmental and health sciences. Lewis Publishers, Boca Raton, FL, USA, pp 189-
523
231.
524
[7] Mackay D, Fraser A. 2000. Bioaccumulation of persistent organic chemicals: mechanisms
525
and models. Environ Pollut 110:375-391.
526
[8] Gobas F, de Wolf W, Burkhard LP, Verbruggen E, Plotzke K. 2009. Revisiting
527
Bioaccumulation Criteria for POPs and PBT Assessments. Integr Environ Assess Manag 5:624-
528
637.
529
[9] Wu JP, Guan YT, Zhang Y, Luo XJ, Zhi H, Chen SJ, Mai BX. 2011. Several current-use,
530
non-PBDE brominated flame retardants are highly bioaccumulative: Evidence from field
531
determined bioaccumulation factors. Environment International 37:210-215.
532
[10] Wu JP, Luo XJ, Zhang Y, Luo Y, Chen SJ, Mai BX, Yang ZY. 2008. Bioaccumulation of
533
polybrominated diphenyl ethers (PBDEs) and polychlorinated biphenyls (PCBs) in wild aquatic
534
species from an electronic waste (e-waste) recycling site in South China. Environ Int 34:1109-
535
1113.
536
[11] Clark KE, Gobas F, Mackay D. 1990. Model of organic-chemical uptake and cearance by
537
fish from food and water. Environ Sci Technol 24:1203-1213.
538
[12] Connell DW. 1989. Biomagnification by aquatic organisms - a proposal. Chemosphere
539
19:1573-1584.
540
[13] Paterson G, Liu JA, Haffner GD, Drouillard KG. 2010. Contribution of fecal egestion to
541
the whole body elimination of polychlorinated biphenyls by Japanese Koi (Cyprinus carpio).
542
Environ Sci Technol 44:5769-5774.
543
[14] Ruus A, Skaare JU, Ingebrigtsen K. 2001. Disposition and depuration of lindane (gamma-
544
HCH) and polychlorinated biphenyl-110 (2,3,3 ',4 ',6-pentachlorobiphenyl) in cod (Gadus
545
morhua) and bullrout (Myoxocephalus scorpius) after single oral exposures. Environ Toxicol
546
Chem 20:2377-2382.
547
[15] Green NW, Schøyen M, Øxnevad S, Ruus A, Høgåsen T, Beylich B, Håvardstun J, Rogne
548
ÅKG, Tveiten L. 2010. Hazardous substances in fjords and coastal waters - 2009. Levels trends
549
and effects. Long-term monitoring of environmental quality in Norwegian coastal waters. TA-
550
2716/2010. Norwegian Climate and Pollution Agency, Oslo, Norway.
551
[16] EC. 2006. Regulation (EC) No 1907/2006 of the European Parliament and of the Council
552
of 18 December 2006 concerning the Registration, Evaluation, Authorisation and Restriction of
553
Chemicals (REACH), establishing a European Chemicals Agency, amending Directive
554
1999/45/EC and repealing Council Regulation (EEC) No 793/93 and Commission Regulation
555
(EC) No 1488/94 as well as Council Directive 76/769/EEC and Commission Directives
556
91/155/EEC, 93/67/EEC, 93/105/EC and 2000/21/EC.
557
[17] Weisbrod AV, Woodburn KB, Koelmans AA, Parkerton TF, McElroy AE, Borgå K.
558
2009. Evaluation of Bioaccumulation Using In Vivo Laboratory and Field Studies. Integr
559
Environ Assess Manag 5:598-623.
560
[18] Knight-Jones E, Knight-Jones P, Nelson-Smith A. 1995. Annelids (Phylum Annelida). In
561
Hayward P, Ryland J, eds, Handbook of the marine fauna of North-West Europe. Oxford
562
University Press, Oxford, UK, pp 165-277.
563
[19] Ruus A, Schaanning M, Oxnevad S, Hylland K. 2005. Experimental results on
564
bioaccumulation of metals and organic contaminants from marine sediments. Aquat Toxicol
565
72:273-292.
566
[20] Lee H, Boese BL, Pelletier J, Winsor M, Specht DT, Randall RC. 1991. Guidance
567
manual: bedded sediment bioaccumulation tests. EPA/600/x-89/302, U.S. Environmental
568
Protection Agency, Pacific Ecosystem Branch, Environmental Research Laboratory, New-Port,
569
OR.
570
[21] Rubinstein NI, Gilliam WT, Gregory NR. 1984. Dietary accumulation of PCBs from a
571
contaminated sediment source by a demersal fish (Leiostomus xanthurus). Aquat Toxicol 5:331-
572
342.
573
[22] Bakke T, Kallqvist T, Ruus A, Breedveld GD, Hylland K. 2010. Development of
574
sediment quality criteria in Norway. J Soils Sediments 10:172-178.
575
[23] Ruus A, Boyum O, Grung M, Naes K. 2010. Bioavailability of PAHs in aluminum
576
smelter affected sediments: evaluation through assessment of pore water concentrations and in
577
vivo bioaccumulation. Environ Sci Technol 44:9291-9297.
578
[24] Karickhoff SW, Brown DS, Scott TA. 1979. Sorption of hydrophobic pollutants on
579
natural sediments. Water Res 13:241-248.
580
[25] You J, Landrum PE, Trimble TA, Lydy MJ. 2007. Availability of polychlorinated
581
biphenyls in field-contaminated sediment. Environ Toxicol Chem 26:1940-1948.
582
[26] You J, Landrum PF, Lydy MJ. 2006. Comparison of chemical approaches for assessing
583
bioavailability of sediment-associated contaminants. Environ Sci Technol 40:6348-6353.
584
[27] Maruya KA, Lee RE. 1998. Biota-sediment accumulation and trophic transfer factors for
585
extremely hydrophobic polychlorinated biphenyls. Environ Toxicol Chem 17:2463-2469.
586
[28] Cornelissen G, Breedveld GD, Kalaitzidis S, Christanis K, Kibsgaard A, Oen AMP. 2006.
587
Strong sorption of native PAHs to pyrogenic and unburned carbonaceous geosorbents in
588
sediments. Environ Sci Technol 40:1197-1203.
589
[29] Beyer A, Wania F, Gouin T, Mackay D, Matthies M. 2002. Selecting internally consistent
590
physicochemical properties of organic compounds. Environ Toxicol Chem 21:941-953.
591
[30] Ellgehausen H, Guth JA, Esser HO. 1980. Factors determining the bioaccumulation
592
potential of pesticides in the individual compartments of aquatic food-chains. Ecotox Environ
593
Safe 4:134-157.
594
[31] Hawker DW, Connell DW. 1985. Relationships between partition-coefficient, uptake rate-
595
constant, clearance rate-constant and time to equilibrium for bioaccumulation. Chemosphere
596
14:1205-1219.
597
[32] Kelly BC, Gobas F, McLachlan MS. 2004. Intestinal absorption and biomagnification of
598
organic contaminants in fish, wildlife, and humans. Environ Toxicol Chem 23:2324-2336.
599
[33] Gobas F, Muir DCG, Mackay D. 1988. Dynamics of dietary bioaccumulation and fecal
600
elimination of hydrophobic organic-chemicals in fish. Chemosphere 17:943-962.
601
[34] Gobas F, McCorquodale JR, Haffner GD. 1993. Intestinal-absorption and
602
biomagnification of organochlorines. Environ Toxicol Chem 12:567-576.
603
[35] Gobas F, Zhang X, Wells R. 1993. Gastrointestinal magnification - the mechanism of
604
biomagnification and food-chain accumulation of organic-chemicals. Environ Sci Technol
605
27:2855-2863.
606
[36] Bruggeman WA, Martron L, Kooiman D, Hutzinger O. 1981. Accumulation and
607
elimination kinetics of dichlorobiphenyls, trichlorobiphenyls and tetrachlorobiphenyls by
608
goldfish after dietary and aqueous exposure. Chemosphere 10:811-832.
609
[37] Sijm D, Seinen W, Opperhulzen A. 1992. Life-cycle biomagnification study in fish.
610
Environ Sci Technol 26:2162-2174.
611
[38] Demain DK, Gallego A, Jaworski A, Priede IG, Jones EG. 2011. Diet and feeding niches
612
of juvenile Gadus morhua, Melanogrammus aeglefinus and Merlangius merlangus during the
613
settlement transition in the northern North Sea. J Fish Biol 79:89-111.
614
[39] Arp HPH, Villers F, Lepland A, Kalaitzidis S, Christanis K, Oen AMP, Breedveld GD,
615
Cornelissen G. 2011. Influence of historical industrial epochs on pore water and partitioning
616
profiles of polycyclic aromatic hydrocarbons and polychlorinated biphenyls in Oslo harbor,
617
Norway, sediment cores. Environ Toxicol Chem 30:843-851.
618
[40] Breivik K, Bjerkeng B, Wania F, Helland A, Magnusson J. 2004. Modeling the fate of
619
polychlorinated biphenyls in the inner Oslofjord, Norway. Environ Toxicol Chem 23:2386-2395.
620 621 622
Figure Legends
623 624
Figure 1. Lipid content (% wet wt.) in liver of cod (Gadus morhua) from the sediment
625
resuspension experiment (left) and the dietary exposure experiment (right) after 13, 26, 39, 52,
626
66, 97 and 129 days; n=3 at all sample days (and both groups; exposed vs. control), except at d 0,
627
where n=6. Median, minimum and maximum are depicted (i.e. all observations, except at d 0). In
628
the sediment resuspension experiment, the ‘exposed’ fish were experimentally exposed to
629
resuspended sediment from the inner Oslofjord, while the ‘control’ fish were not exposed to
630
sediment. In the dietary exposure experiment, the ‘exposed’ fish were fed polychaetes (Nereis
631
virens) previously exposed to sediment from the inner Oslofjord, while the ‘control’ fish were fed
632
N. virens previously exposed to unpolluted sediment. Note: Categorical X-axis.
633 634
Figure 2. Concentrations (µg kg-1; lipid wt.) of PCBs (-28 , -52, -101, -118, -153, -138 and -180,
635
and the sum of these, PCB7) in liver of cod (Gadus morhua) from the sediment resuspension
636
experiment after 13, 26, 39, 52, 66, 97 and 129 days; n=3 at all sample days (and both groups;
637
exposed vs. control), except at d 0, where n=6. Median, minimum and maximum are depicted
638
(i.e. all observations, except at d 0). The ‘exposed’ fish were experimentally exposed to
639
resuspended sediment from the inner Oslofjord, while the ‘control’ fish were not exposed to
640
sediment. Significant differences between ‘exposed’ and ‘control’ are indicated by “*”.
641
Significant differences among sampling days in the exposed group are indicated by “a”, while
642
significant differences among sampling days in the control group are indicated by “b”.
643
Significant differences between each specific sampling day and d 0 are indicated by “c”. Note:
644
different scale on response axes; categorical X-axis.
645
646
Figure 3. Concentrations (µg kg-1; lipid wt.) of PCBs (-28 , -52, -101, -118, -153, -138 and -180,
647
and the sum of these, PCB7) in liver of cod (Gadus morhua) from the dietary exposure
648
experiment after 13, 26, 39, 52, 66, 97 and 129 days; n=3 at all sample days (and both groups;
649
exposed vs. control), except at d 0, where n=6. Median, minimum and maximum are depicted
650
(i.e. all observations, except at d 0). The ‘exposed’ fish were fed polychaetes (Nereis virens)
651
previously exposed to sediment from the inner Oslofjord, while the ‘control’ fish were fed N.
652
virens previously exposed to unpolluted sediment. Significant differences between ‘exposed’ and
653
‘control’ are indicated by “*”. Significant differences among sampling days in the exposed group
654
are indicated by “a”, while significant differences among sampling days in the control group are
655
indicated by “b”. Significant differences between each specific sampling day and d 0 are
656
indicated by “c”. Note: different scale on response axes; categorical X-axis.
657 658
Exposed Control
Dietary exposure experiment Sediment resuspension experiment
Day
Lipids (% w. wt.)
0 13 26 39 52 66 97 129
0 20 40 60 80 100
Day
Lipids (% w. wt.)
0 13 26 39 52 66 97 129
0 20 40 60 80 100
